what are our imaging tools? x-ray based imaging...milieu what are our imaging tools? • x-ray based...

1

Measurement Issues in Imaging

Richard L. Wahl, M.D.

Division of Nuclear Medicine

Johns Hopkins Medical Institutes

Baltimore, MD

Disclosure InformationINTRODUCTION TO CLINICAL RESEARCH: A TWO-WEEK

INTENSIVE COURSE, 2010Richard L. Wahl, M.D.

I have the following financial relationships to disclose:Consultant for: Cellectar, Nihon MediPhysics, MDS Nordion; GSKGrant/Research support from: NIHStockholder in: Threshold PharmaceuticalsHonoraria from: GSKEmployee of: Johns Hopkins UniversityRoyalties and Patents licensed to: Glaxo Smith Kline, Spectrum Pharmaceuticals, Naviscan PET; Intramedical Imaging

.

- and -

- or -

I will discuss the following off label use and/or investigational use in my presentation: FLT (PET Agents)

Why Include Imaging in a Clinical Trial?

Primary evaluation of the imaging test for:Diagnostic accuracyPrognostic accuracyPredictive accuracyAs a biomarker

Evaluation as an endpoint for a therapy:for example, response or non response to a cancer therapy

Amyloid burden in the brain for a trial of an AD drug

Secondary correlative marker. Eg. Survival a primary endpoint, tumor shrinkage a secondary endpoint.

Why Include Imaging in a Trial?

• Varying pharmacokinetics in whole body, organs and tumors

• Varying receptor status in tumors• Varying proliferative/apoptotic rates in tumors• Varying in many characteristics• Varying in many characteristics• Current cancer treatments are substantially based

on the “Average Patient” not the individual patient.

• Goal is to individualize treatments to optimize responses and minimize toxicity.

• Imaging shows the phenotype in the physiological milieu

What are our Imaging Tools?

• X-ray based imaging: Chest X ray, Bone radiograph, CT scan, angiogram

• Sound based—Ultrasound (including Doppler US)

• Nuclear Medicine—Gamma camera imaging,Nuclear Medicine Gamma camera imaging, PET, SPECT imaging

• Magnetic Resonance Imaging-- including varying pulse sequences, DCE MRI,

• Optical Imaging: Human eye, bioluminescence, NIR (best in animals)

X-Ray Based Imaging• Strengths:

– Excellent anatomic resolution– Quick and widely available– Provides size measurement used in RECIST

Weakness:Weakness:Ionizing radiationInformation content somewhat limitedTumor size changes are “last thing” to occur with successful or unsuccessful rx.Lack of specific contrast agents

2

FDG-PET/CT imaging (FOV 25 cm)

relative tumor volume by CT (Observer2)

200

250

0

50

100

150

0 1 2 3

days after chemo

%

treated

control

Ultrasound Based Imaging• Strengths

– No ionizing radiation– Reasonably good resolution– Relatively Inexpensive– Widely availableWidely available

– WeaknessesLimited depth of penetrationLimited quantitative informationLack of specific contrast

Nuclear MedicinePET and Single Photon Imaging

• Strengths– Low mass quantities required (tracer)– Quantiative– Can evaluate deep structures– Wide range of tracers available (or not)

Can combine with CT– Can combine with CT– Reproducibility understood

• Weaknesses– Limited resolution– Ionizing radiation– Radiotracers may not be widely available– Cost– Short half lifes make logistically challenging

Magnetic Resonance Imaging

• Strengths– Relatively widely available– Resolution is good and can be excellent– Many pulse sequences available– Can provide physiological information

Some contrast agents emerging (nodal/vascular)– Some contrast agents emerging (nodal/vascular)• Weaknesses

– Predominantly anatomic in many applications– Pulse sequences are not well-standardized across

manufacturers or institutions– Reproducibility is not yet well characterized– Cost– NSF (severe complication in patients with poor renal

function and repeated Gadolinium dosing)

3

Breast Applications

Baseline Before surgery

M

S

PAL

I

1st cycle (7

days)

4th cycle (5

days)

Jacobs MA, Ouwerkerk R, Wolff AE, Stearns V, Bhujwalla, ZM , Bluemke D.

Optical Imaging

• Strengths– Very sensitive to small numbers of molecules

– Semi quantitative (attenuation effects)

– Some use in OR and superficial organ assessments

– Excellent pre-clinical tool

• Weaknesses– Not easily scalable to humans, at least for deep

structures due to light absorption

– Limited specific contrast agents (but they can be made)

Results

d0d4d8d11d18

Treated (35 uCi/mouse) Untreated

Is imaging tool “Fit for the Task?”

• Must understand question one wishes to address

• Must understand limitations and strengthsMust understand limitations and strengths of imaging tool

• Consultation with a subject expert is recommended if imaging has a major role in a clinical trial.

Qualitative Analysis vs Quantitative

• Variability in the Interpretation of Screening

• Mammograms by US RadiologistsMammograms by US Radiologists

• Findings From a National Sample

• Craig A. Beam, PhD; Peter M. Layde, MD, MSc; Daniel C. Sullivan, MD

• Arch Int. Med 1996; 156:209-213

Qualitative reads are Qualitative

• Not all readers are created equally

• Performance can and does vary among readersreaders

• For test/re test studies, it seems apparent that the same reader should be used for both the baseline and the follow up

4

Results of Multi Reader Study• There is a range of at least 40% among US

radiologists in their screening sensitivity.

• There is a range of at least 45% in the rates at which women without breast cancer arewhich women without breast cancer are recommended for biopsy.

• As indicated by receiver operating characteristic curve areas, the ability of radiologists to detect cancer mammograms varies by as much as 11%.

Biases in Imaging

• Reader experience

• Variation in equipment or methods in a center or across centerscenter or across centers

• Reader blinding to truth data

• Stage migration

• Verification bias

• Dichotomous of continuous recording of results?

B+

Detector 2

Positron Emitter

Positron Emission,Annihilation, Detection

B+B+B-

Detector 1

Emitter

Coincidence CircuitryComputer

Images

PET/CT Design

5

Structure and Affinity of the Neutral Thioflavin-T Analogue [11C]PIB

S

N

HO 11CH3S

N

CH3 CH3

N

NH

[11C]6-OH-BTA-1 or [11C]PIB

N

NCH3

thioflavin-TCH3

+

A(1-40) Ki = 580 nM A(1-40) Kd = 2 nM

[18F] Av-45Collaboration JHU Radiology, Psychiatry and

AVID pharma Nov 2007

2727

Top = Healthy Control 83 year old MaleBottom = AD 78 year old Male

Impact of PET/CT in Cardiac Disease

• Single photon imaging is established throughout the world as a tool to individualize patient treatments

• While very effective, is less accurate in RCA distribution and in larger patients such as in the US

• Limitations for viability• Limitations for “balanced” disease• PET/CT and PET are more accurate in

many studies. When should they be used? All/some/No patients?

Cardiac Imaging: Biological Considerations

• Myocardium:– Flow

– ViabilityViability

– Innervation

– Necrosis/Apoptosis

– Blood vessels: plaque, inflammation

Ru-82 PET Stress/Rest Polar Map

6

Dynamic Rb-82 PET-CT for Quantification of Myocardial Blood Flow

Rb Flow vs Microsphere

CC

Dog Model of Myocardial Ischemia Lautamäki et al, EJNMMI 20

Cardiac PET/CT

• Obtain Myocardial Perfusion images (PET)

• Obtain Structural and anatomical abnormalities within the field of view (CT)abnormalities within the field of view (CT)

• Obtain Calcium score (CT)

• Obtain Coronary structure and detection of lesions (CTA)

• Obtain area(s) of viable tissue (PET)

Efficacy of Current Cancer Rx• 100% of patients get Rx.

• 10-15% in metastatic lung, esophagus, pancreatic cancers

• Paradigm of treating 10 to benefit 1 is notParadigm of treating 10 to benefit 1 is not economically sustainable

• Need PERSONALIZED approaches to enrich response rates

• Benefit to those who will benefit most and avoid harm to those who will not

PET/CT and SPECT/CT are Quantitative Methods

• Most applications to date have been Qualitative

• In treatment response assessment, especially if looking for small treatment induced changes, Quantitation will be needed

• Quantitation requires greater attention to technical details than qualitative imaging

• Standardization of methods is required

Potential Roles of Advanced Imaging in Drug Development

• Pre-Clinical– In vitro– Animal Models

• Clinical– Phase I– Phase II– Phase III

Imaging What?

• Drug Itself?• Surrogate for the drug? (eg. In-111 for Y90)• Tumor characteristics best suited to drugTumor characteristics best suited to drug

– Receptor or hypoxia, proliferation rate, blood flow

• Tumor response to the drug– FDG, FLT, size, vascular permeability or flow

• Size:

7

Roles of PET/CT and SPECT/CT in Cancer Management

• Prior to treatment– Diagnosis (cancer or not?)—including directing biopsy– Staging before therapy

• Pre surgical• Pre Radiation therapy (planning)

• Pre Radionuclide therapy (dosimetry)– Predicting Response*

• After treatment initiated– Assessment of response after completing therapy– Intra-therapy monitoring– Response Adaptive Approaches

Imaging What?



• Tumor response to the drug– FDG, FLT, size, vascular permeability or flow

• Size:

Therapeutic Dose450 mg unlabeled tositumomab

Day 7–14

Thyroid protective agent: Day –1 continuing through 14 days post-therapeutic dose

I-131 Tositumomab Treatment Regimen

Day 0

Dosimetric Dose450 mg unlabeled tositumomab Total Body 450 mg unlabeled tositumomab,

35 mg tositumomab radiolabeled I 131 to deliver specific cGy TBD (variable mCi)

• Unlabeled dose infused over 1 hour

• Radiolabeled therapeutic dose infused over 20 minutes

450 mg unlabeled tositumomab,35 mg tositumomab

radiolabeled I 131 (5 mCi)

• Unlabeled dose infused over 1 hour

• Radiolabeled tracer dose infused over 20 minutes

Total Body Counts x 3

• Day 0• Day 2, 3, or 4• Day 6 or 7

8

Imaging What?



• Tumor response to the drug– Size: RECIST 1.0, 1.1; WHO, Volume– FDG, FLT, size, vascular permeability or flow

Days 1 15 27

Rituximab (375 mg/m2)

Weekly ( 4)

Stem cell collection ACST*

90Y ibritumomab tiuxetan

(dosed by cohort*)

111In ibritumomab tiuxetan (5 mCi)

Cyclophosphamide (2.5 g/m2)

Single dose

Weeks

Imaging dose Therapeutic dose

Organ Dosimetry Dose Escalation Study of 90Y Ibritumomab Tiuxetan and ASCT: Study Schema

Days 1 15 ~27

Filgrastim (10 g/kg)

Dosimetry by planar and

SPECT/CT scans

Weeks

• 0.4 mCi/kg

• 14–18 Gy to the liver

• 24 Gy to the liver

• 28 Gy to the liver

*Dose cohorts of 90Y ibritumomab tiuxetan

Range of mCi/kg/10Gy

mCi/kg/10Gy

0 8

1

ds

0

0.2

0.4

0.6

0.8

0 5 10 15 20

patient number

mC

i/kg

/10

00

ra

Series1

.43-.86 mCi/Kg range for delivery of 10Gy to target organ

66

46 4660

80

100

CR

PRPatients (%) 10/15

20

813

0

20

40

60 days aftertransplantation

6-month follow-up

(%) 10/15

2/153/15

6/13

1/13

Progressive disease

6/13

Roles of PET/CT in Cancer Management


• Pre surgical• Pre Radiation therapy (planning)• Pre Radionuclide therapy (dosimetry)

– Predicting Response*• After treatment initiated

– Assessment of response after completing therapy– Intra-therapy monitoring– Response Adaptive Approaches

9

Imaging What?



• Tumor response to the drug– Size: RECIST 1.0, 1.1; WHO, Volume– FDG, FLT, size, vascular permeability or flow

Anatomic Response II

Figure 1. Changes in response after the WHO and RECIST criteria were separately applied to the same patients.

Park J O et al. Jpn. J. Clin. Oncol. 2003;33:533-537

©2003 by Oxford University Press

Limitations of Anatomic Response

• Slow• Not sensitive to early response in some

tumor types (GIST, Sarcoma)yp ( )• Variability in measurement• Tumor Size a late downstream marker• Difficult to detect some tumors• Hard to separate tumor from treatment

response or other pathology

10

Molecular and Functional Alterations in Cancer I.

• Increased glucose metabolism

• Increased Amino acid transport

• Increased Protein and membrane synthesis• Increased Protein and membrane synthesis

• Increased DNA synthesis

• Overexpression of receptors/antigens

• Increased blood flow or vessel density

• Decreased oxygen tension in lesions

FDG FDG FDG-6-PK1 K3

FDG: A Tracer of EarlySteps of Glucose Metabolism

G G G 6K2 K4

Extracellular Intracellular Intracellular

FDG Signal

• “Heartbeat of the Cell”

• Downstream Marker of Many Processes

Why do we need PET in Treatment Response Assessment?

• PET can detect tumors not detectable by CT

• PET responses often occur earlier than those seen on CTthose seen on CT

• PET has higher predictive ability at end of treatment than CT

• PET may be more reproducible than CT for measuring tumor response to therapy

11

Post-therapy PET• Multiple studies demonstrating superior prognostic

value of PET vs CT at therapy completion– Can differentiate viable tumor from fibrosis

– Can detect residual disease in context of radiographic CR

• Relapse up to 100% if PET+, 5-17% if PET-

Spaepen K et al, JCO 2001; 19: 414-419

P < 0.0001

Jerusalem et al, Blood 1999; 94:429-433

Bone Marrow Involvement

12

IWG response criteria

Original Revised (includes Hodgkin’s)

CR Normal size FDG-avid tumor: mass of any size, as long as PET “negative”

Variably FDG-avid or unknown: normal size

CRu If nodes > 1.5 cm, > 75% decrease

n/a

PR > 50% decrease FDG-avid tumor: > 50% decrease, but at least one PET positive focus

Variably FDG-avid or unknown: regression in size

Cheson BD et al, JCO 2007; 25: 579-586

“International Harmonization Project (IHP) Recommendations for FDG-PET in Patients

with Lymphoma (abst 24, EANM, Juweid et al)

– Recommendations for PET, not PET-CT– PET mandatory for post Rx HD and

DLBCL (pre Rx PET recommended)– PET pre Rx strongly recommended forPET pre Rx strongly recommended for

variably FDG avid NHL and post Rx (if clinical trial)

– “Mid Treatment” PET only should be done in Clinical Trial

– “No justification” for routine post-Rx surveillance with PET.





– Assessment of response after completing therapy

– Intra-therapy monitoring– Response Adaptive Approaches

Baseline 5/03 2 cycles chemo 6/03 4 cycles chemo 7/03

13

When is Response Assessed?

• End of a Treatment Course: perhaps most relevant to have a complete response at this time

Mid treatment course

Shortly after treatment is initiated

Why might mid-treatment PET be superior to post-treatment?

Early PET result implies a certain rate of tumor kill

1 E+051.E+061.E+071.E+081.E+091.E+101.E+111.E+12

Usual size at diagnosis

om

a ce

lls

First-order kinetics

With 6 cycles, need at least 1.5 logs of cell kill per cycle

1.E+001.E+011.E+021.E+031.E+041.E+05

0 1 2 3 4 5 6

Cure

Cycles of chemotherapy

Logs

of

lym

pho 1 E+05

1.E+061.E+071.E+081.E+091.E+101.E+111.E+12

Usual size at diagnosiso

ma

cells

Lower detection limit of PET PET likely can only

measure the first 2-3 logs of cell kill (so negative PET does not mean b f t )

1.E+001.E+011.E+021.E+031.E+041.E+05

0 1 2 3 4 5 6

Cure


Logs

of

lym

pho absence of tumor)

1 E+051.E+061.E+071.E+081.E+091.E+101.E+111.E+12


om

a ce

lls

Lower detection limit of PET

A true negative PET after 2 cycles implies an adequate rate of tumor kill

A true negative PET at end of therapy might be less predictive

1.E+001.E+011.E+021.E+031.E+041.E+05

0 1 2 3 4 5 6

Cure


Logs

of

lym

pho p 1 E+05

1.E+061.E+071.E+081.E+091.E+101.E+111.E+12


om

a ce

lls


1.E+001.E+011.E+021.E+031.E+041.E+05

0 1 2 3 4 5 6

Cure


Logs

of

lym

pho

14

1 E+051.E+061.E+071.E+081.E+091.E+101.E+111.E+12


om

a ce

lls


A true positive PET after 2 cycles suggests cure is unlikely

1.E+001.E+011.E+021.E+031.E+041.E+05

0 1 2 3 4 5 6

Cure


Logs

of

lym

pho

Practical Considerations:

• Consistent Protocol Essential

• Same time from injection to imaging

• Same fasting• Same fasting

• Same Imaging sequence (caudal to cranial)

• Variability in estimates of SUV of 10-20%

• Changes in SUV under 20% may be chance

• Normalization to liver can help

Mid-treatment PET strongly predicts PFS in aggressive lymphoma

p < 0.00001 PET + 1.5 mo

Median time to treatment failure

p < 0.00001

PET - 1 yr

Spaepen K et al, Ann Oncol 2002;13: 1356-63

Stronger predictor than IPI





– Assessment of response after completing therapy– Intra-therapy monitoring

– Response Adaptive Therapy

Response-adapted therapy

J0348: rationale• Early [18F] FDG-PET (after 2-4 cycles) is highly predictive

and more individualized than the IPI• Event rates of 78-100% if PET+ midtreatment,

versus 8-16% if PET-, consistently seen

• Benefit of early BMT most pronounced in poor-risk ptsBenefit of early BMT most pronounced in poor risk pts

• HYPOTHESIS: Early PET can identify those pts who most stand to benefit from early treatment intensification

• PRIMARY ENDPOINT: 25% absolute improvement in 2-yr EFS (from 20% to 45%) of PET+ pts, compared with historical outcomes, through early BMT

15

Aggressive NHL, any stage, any IPI

PET +PET

(R)CHOP for 2 or 3 cycles

Study design

PET +PET -

complete conventional therapy

if no disease progression

(R)ESHAP or (R)ICE x 2

High dose therapy and ABMT

J0348

• Primary endpoint: 25% absolute improvement in 2-yr EFS (from 20% to 45%) of PET+ pts, compared with historical outcomes through early BMThistorical outcomes, through early BMT

• Goal: 55 pts, expect 50% PET + and 19 transplants

PET assessment

NEGATIVE• 0 no abnormal activity (tumor cold compared with

background)• 1+ minimal activity (tumor less than background)• 2+ equivocal (tumor = background)• 2+ equivocal (tumor = background)_____________________________________________

POSITIVE• 3+ moderate activity (tumor greater than background)• 4+ strong activity (tumor much greater than

background)

Response Adaptive Therapy• Fifty-nine newly diagnosed patients, 98% with B cell lymphoma, had

PET/CT performed after 2 or3 cycles of first-line chemotherapy.

• Mid-treatment PET was positive in 33 (56%);

• 28 received ASCT with an actuarial 2-year EFS of 75% (95% confidence interval 60% 93%)confidence interval, 60%-93%).

• On intention-to treat analysis, 2-year EFS was 67% (53%-86%) in all PET-positive patients and 89% (77%-100%) in PET-negative patients. (overlapping CI)

• Kasamon, Wahl, Ziessman et al. BBMT 15:242-248, 2009

E3404: Phase II Study of Response-Adapted Therapy

Newly diagnosed large B-cell lymphoma

R-CHOP x 4

PET + R-ICE X 4

lymphoma

(Stage III, IV, or Bulky II)

PET during C3

PET - R-CHOP X 2

Central review of mid-treatment PET, designated + or -

Conclusions:

• PET/CT after initial therapy is complete has prognostic value in lymphomas > CT alone.

• PET/CT after 1-2 cycles appears predictive of effectiveness of treatment in several tumor types. Will allow for response adaptive therapy.

• PET/CT with non FDG ligands, eg. FES, may be predictive before treatment is begun.

16

FLT Lung Cancer Response

• FLT-PET images of the thorax were obtained before and 7 days after the start of gefitinib (250 mg/d) therapy in non smokers with new or recurrent NSCLC

• SUV max and % decline in SUV max were assessed in tumor at baseline and 7 days Compared with CT after 6 weeks ofbaseline and 7 days. Compared with CT after 6 weeks of therapy.

• 31 patients of whom 28 had complete data.

• CT at 6 weeks showed PR in 14 (50%), SD in 4 (14%), and PD in 10 (36%) after 6 weeks of treatment.

• Pretreatment SUVmax of the tumors did not differ between responders and nonresponders.

Sohn et al. Clin Cancer Res. 2008 Nov 15;14(22):7423-9.

Fig. 2

Copyright ©2008 American Association for Cancer Research

Sohn, H.-J. et al. Clin Cancer Res 2008;14:7423-7429

•At 7 days after the initiation of therapy, the percent changes in SUVmax were significantly different (-36.0 +/- 15.4% versus 10.1 +/-19.5%; P < 0.001). Decrease of > 10.9% in SUVmax was used as the criterion for predicting response.



Fig. 4



Fig. 3



Qualitative vs Quantitative or Both?

Visual assessments can be used to assess response

Strengths: No special instrumentation, integration of all data. W k T d t b bi t Weaknesses: Tendency to be binary, not reproducible in several series

Quantitative: Ignores qualitative data, can be erroneous due to technical factors, “Standardized” is not so standard after all

17

Practical Considerations:

Consistent Protocol Essential

Same time from injection to imaging

Same fasting Same fasting

Same Imaging sequence (caudal to cranial)

Variability in estimates of SUV of 10-20%

Changes in SUV under 20% may be chance

EORTC 1999

EORTC Response CriteriaCR Complete Disappearance of all Metabolically

Active Tumor (i.e. decreased to background levels)

PR >15% decline in SUV after 1 cycle, >25% decline after 2 or more cycles. Reduction in yextent (size) of FDG uptake is not required

SD Increase in FDG SUV of <25% or decrease of <15% in SUV and no increase in extent of uptake (<20% in longest dimension)

PMD Increase in SUV of over 25%, Increase in extent of FDG uptake by >20%, New FDG positive metastases

If there were RECIST Criteria for PET, It Would be Defined

as…

“PERCIST”

• Positron

• Emission

• Response• Response

• Criteria in

• Solid

• Tumors

18

J Nucl Med 2009 50: 122S-150S

Figure 1a. Graphs depict the relationships between patient body weight and blood SUVs: (a) SUVbw, (b) SUVibw, (c) SUVlbm, or (d) SUVbsa

Copyright ©Radiological Society of North America, 1999

Sugawara, Y. et al. Radiology 1999;213:521-525

Figure 1c. Graphs depict the relationships between patient body weight and blood SUVs: (a) SUVbw, (b) SUVibw, (c) SUVlbm, or (d) SUVbsa

Copyright ©Radiological Society of North America, 1999

Sugawara, Y. et al. Radiology 1999;213:521-525

What is Measured?• SUL Peak

• In Hottest Tumor Focus

What is Measured?• PERCIST 1.01. Measurable target lesion is hottest single tumor lesion SUL of

“maximal 1.2 cm diameter volume ROI in tumor” (SUL peak). SUL peak is at least 1.5 fold greater than liver SUL mean+ 2SD(in 3 cm spherical ROI in normal right lobe of liver).

• If liver is abnormal, primary tumor should have uptake >2.0 times SUL mean of blood pool in a 1 cm diameter ROI in the descending thoracic aorta extended over 2 cm Z axis.

Th i h h i SUL k i d Whil i ll hi• The tumor with the maximum SUL peak is assessed post -treatment. While typically this is in the same region of the tumor with the highest SUL peak at baseline , it need not be.

• Uptake measurements should be made for the peak and maximum single voxel tumor SUL.

• Other SUV metrics including SUL mean at 50 or 70% of SUV peak can be collected as exploratory data, TLG can be collected ideally based on voxels more intense than 2SD above liver mean SUL (see below). 4. These parameters on up to 5 measurable target lesions can be recorded as exploratory data, typically the 5 hottest lesions, which are typically the largest, not over 2 /organ. Tumor size of these lesions can be determined per RECIST 1.1.

19

Introduction• A number of different ROI

definitions have been employed including:

Isocontour

– Mean within an irregular ROI defined by isocontours.

– Mean within a fixed size ROI centered on the most metabolically active region.

Maximum pixel

Fixed size

Results: Statistical Quality

3D

2D

• Statistical quality of the images deteriorates with decreasing scan duration for both 2D and

0.5 min 2 min 3.5 min 5 min 20 min

3D

Results: ROImaxRecovery coefficients for the 25 mm diameter insert determined using ROImax.

• ROImax increasingly overestimates recovery as noise increases (shorter scan durations).

Results: ROImax0.5 minute acquisition

SUVmax = 5.28

20 minute acquisition

SUVmax = 2.57

• Insert has an SUV of 2.5 (2.5:1 insert-to-background ratio)

20

Normalization for Quality Control• -Normal liver SUL must be within 20% (and <0.3 SUL mean units) for

baseline and follow up study to be assessable.-If liver is abnormal, blood pool SUL must be within 20% (and <0.3 SUL mean units) for baseline and follow up study to be assessable.-

• Uptake time of baseline study and follow up study 2 must be within 15 minutes of one another to be assessable. Typically, these are at a mean yp y,of 60 minutes post injection, but not less than 50 min post injection. –

• Same scanner, or same scanner model at same site, injected dose, acquisition protocol (2D vs. 3D), and software for reconstruction should be used. - Scanners should provide reproducible data and be properly calibrated.

Complete Metabolic Response

• Complete metabolic response (CMR) complete resolution of [18F]-FDG uptake within the measurable target lesion so that it is less than mean liver activity and indistinguishable from surrounding background blood pool l llevels .

• Disappearance of all other lesions to background blood pool levels . % decline in SUL should be recorded from measurable region as well as (ideally) time in weeks after treatment was begun (i.e. CMR -90, 4).

• No new FDG avid lesions in a pattern typical of cancer. If progression by RECIST must verify with follow up

A

B

Continued declines out to 24 weeks

C

CT PET FUSED

Partial metabolic response (PMR)

• Reduction of a minimum of 30% in target measurable tumor FDG SUL peak.

• Absolute drop in SUL must be at least 0.8 SUL units, as well.

• Measurement is commonly in the same lesion as the baseline, but can be another lesion if that lesion was previously present and is most active lesion after treatment. ROI does not have to be in precisely the same area as baseline scan, though typically it is.

• No increase, >30% in SUL or size of target or non target lesions (i.e. no PD by RECIST or IWC) (.If PD anatomically, must verify with follow up). A reduction in the extent of the tumor FDG uptake is not a requirement for partial metabolic response. % decline in SUL should be recorded as well as (ideally) time in weeks after treatment was begun (i.e. PMR -40, 3). No new lesions.

1

2

21

Stable Metabolic Disease

• Stable metabolic disease (SMD) Not CMR, PMR nor PMD. Note, the SUL peak in metabolic target lesion should be grecorded as well as (ideally) time from start of most recent therapy in weeks (i.e. SMD -15,7). No new lesions

Progressive metabolic disease (PMD)

• >30% increase in FDG SUL peak, with >0.8 SUL unit increase in tumor SUV peak from the baseline scan in pattern typical of tumor and not of infection/treatment effect.

• OR- Visible increase in the extent of [18F]-FDG tumor uptake (75% in TLG l ith d li i SULTLG volume with no decline in SUL.

• OR - new [18F]-FDG avid lesions which are typical of cancer and not related to treatment effect or infection. PMD other than new visceral lesions should be confirmed on follow up study within 1 month unless

• PMD also is clearly associated with progressive disease by RECIST 1.1. PMD should be reported to include % change in SUV peak, (ideally time post treatment in weeks) and whether new lesions are present/absent and their number (i.ePMD, +35, 4, New-5).

1

2

PERCIST Continous Response Scale• Because SUL is a continuous variable, dividing response

criteria into a limited number of somewhat arbitrary response categories loses much data.

• For this reason PERCIST preserves percent declines in the SUV peak in each reported category Because theSUV peak in each reported category. Because the rapidity with which a scan normalizes is important (faster appears better), PERCIST asks for the time from start of treatment as part of the reporting.

• For example, a CMR 90, 1 is probably superior to a CMR 90, 10, especially if the latter patient were SMD 20,1. More than one measurement of PET response may be needed at differing times and it may be treatment type dependent.

Number of Lesions• PERCIST 1.0 only evaluates the SUL peak of the hottest

tumor. This is a possible limitation of the approach, but lesions and their responses are highly correlated in general.

• Additional data are required to determine how manyAdditional data are required to determine how many lesions should be assessed over 1.

• A suggested option is to include the 5 hottest lesions, or the 5 observed on RECIST 1.1 which are the most measurable. % change in SUL can be reported for the single lesion with the largest increase in uptake or the smallest decline in uptake. Additional studies will be needed to define how many lesions are optimal for assessment.

22

Inter-Observer Variability of SUV

Same tumor data set measured multiple times by independent observers

100% agreement in SUV determination – Minn et al, Radiology, 1995

10 tumors each measured twice by 2 independent observers– 10 tumors each measured twice by 2 independent observers

– Semi-automated image analysis software

“Good” inter-observer agreement – Marom et al J Thorac Imaging 2006

– 5 readers measured 20 primary tumors four times

Untreated primary lung cancers on average > 2 cm

Assessment of Inter-observer Reproducibility in Quantitative

FDG PET and CT Measurements of Tumor Response to Therapy

HA Jacene, S Leboulleux, S Baba, D Chatzifotiadis, B Goudarzi, O Teytelbaum, Horton, I Kamel, K Macura, H

Tsai, J Kowalski and RL Wahl J Nucl Med 2009;50:1760-9.

Objectives

To directly compare inter-observer reproducibility of

1) SUV & CT size measurements in malignant tumors pre and post therapytumors pre- and post-therapy

2) % change in SUV & CT size measurements in response to therapy

Jacene et al. J Nucl Med 2009;50:1760-9

Percent change SUVbw max

ICC – 0.94Jacene et al. J Nucl Med 2009;50:1760-9

Percent change Longest CT size


Percent change 2D CT size


24

Measurement Issues: Summary

• Rigor of imaging approach depends on if imaging is a primary of secondary/tertiary aim

• Rigorous use of imaging requires attention to details including reader selection, methods of acquisition and methods of analysis

• Clinical reads, especially qualitative reads, are highly variant depending on skill sets of readers

• Major national push to quantitative imaging biomarkers applicable across disease types

• Quantitative imaging should add rigor to clinical studies

Acknowledgements• Heather Jacene, M.D.

• Eric Frey, Ph.D.

• Bin He, Ph.D.

• George Sgouros, Ph.D.

• Martin Lodge, Ph.D.

• Wayne Kasecamp, CNMT

• Ken Zasadny, Ph.D.

• James Engles, M.S., MBA.

• Ian Flinn, M.D.

what are our imaging tools? x-ray based imaging...milieu what are our imaging tools? • x-ray based...

Documents